Accuracy of emergency physicians' probability estimates for acute coronary syndrome

Accurate risk assessment is crucial for appropriate management of patients presenting to the emergency department (ED) with chest pain. This study aimed to evaluate how accurately emergency physicians estimate the probability of acute coronary syndrome (ACS).

A vignette-based on a real ED chest pain patient was presented in four sequential frames with increasing clinical information: history and symptoms, ECG findings, and elevated troponin levels. Evidence-based reference probabilities of ACS were established for each frame. Emergency physicians in the Skåne region of Sweden were invited to estimate the probability of ACS at each stage. The primary outcome was the difference between physicians' estimates and the reference probabilities.

Of 707 respondents, 357 physicians with recent ED experience provided complete responses and were included in the analysis. The median age was 34 years (IQR 30-40), 56 % were male, and slightly more than half were specialists or senior residents. Physicians consistently overestimated the probability of ACS across all frames. The largest discrepancy occurred in the initial history/symptoms frame, where the median estimated probability was approximately 70 %, compared with a reference range of 3-16 %. Although estimation accuracy improved as additional diagnostic information was provided, estimates remained higher than reference probabilities throughout.

In this realistic staged case, ED physicians substantially overestimated the risk of ACS in chest pain patients. Such overestimation may contribute to unnecessary diagnostic testing and overtreatment, highlighting the need for improved training and/or decision-support tools in emergency care.